Professional Documents
Culture Documents
Evidence-Based Analysis of Risk Factors For Postoperative
Evidence-Based Analysis of Risk Factors For Postoperative
Background. In assessing a patient’s risk for postoperative nausea and vomiting (PONV), it is
Editor’s key points important to know which risk factors are independent predictors, and which factors are not
† Identifying independent relevant for predicting PONV.
predictors for Methods. We conducted a systematic review of prospective studies (n.500 patients) that
postoperative nausea applied multivariate logistic regression analyses to identify independent predictors of PONV.
and vomiting (PONV) Odds ratios (ORs) of individual studies were pooled to calculate a more accurate overall
would be useful. point estimate for each predictor.
† Systematic review of 22 Results. We identified 22 studies (n¼95 154). Female gender was the strongest patient-
large (.500 patients specific predictor (OR 2.57, 95% confidence interval 2.32 –2.84), followed by the history of
each) studies identifying PONV/motion sickness (2.09, 1.90 –2.29), non-smoking status (1.82, 1.68– 1.98), history
predictors of PONV. of motion sickness (1.77, 1.55–2.04), and age (0.88 per decade, 0.84 –0.92). The use of
† Female, previous PONV, volatile anaesthetics was the strongest anaesthesia-related predictor (1.82, 1.56 –2.13),
non-smoker, younger followed by the duration of anaesthesia (1.46 h21, 1.30– 1.63), postoperative opioid use
age, volatile (1.39, 1.20 –1.60), and nitrous oxide (1.45, 1.06 –1.98). Evidence for the effect of type of
anaesthetics, and surgery is conflicting as reference groups differed widely and funnel plots suggested
postoperative opioids significant publication bias. Evidence for other potential risk factors was insufficient
were predictors. (e.g. preoperative fasting) or negative (e.g. menstrual cycle).
† Some factors commonly Conclusions. The most reliable independent predictors of PONV were female gender, history
thought to be predictors of PONV or motion sickness, non-smoker, younger age, duration of anaesthesia with volatile
were not. anaesthetics, and postoperative opioids. There is no or insufficient evidence for a number of
commonly held factors, such as preoperative fasting, menstrual cycle, and surgery type, and
using these factors may be counterproductive in assessing a patient’s risk for PONV.
Keywords: PONV; risk; vomiting, nausea, anaesthetic factors; vomiting, nausea, patient
factors; vomiting, nausea, surgical factors
Accepted for publication: 24 May 2012
Postoperative nausea and/or vomiting (PONV) is an unpleas- resulting complications are costly for the healthcare sector
ant experience that afflicts 20 –30% of surgical patients after worldwide, with several hundred million dollars spent annu-
general anaesthesia.1 PONV decreases patient comfort and ally in the USA alone.10
satisfaction, and, rarely, may cause dehydration and electro- PONV is a multifactorial phenomenon that can be triggered
lyte imbalances, aspiration of gastric contents, oesophageal by multiple receptor pathways at peripheral, central, or both
rupture, suture dehiscence, and bleeding.2 – 9 PONV and its sites.11 A number of patient-specific, anaesthesia-related,
†
This work was presented as a poster at the American Society of Anesthesiologists Annual Meeting 2009, New Orleans, LA, USA.
‡
Joint first-authorship.
& The Author [2012]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
For Permissions, please email: journals.permissions@oup.com
Systematic review of PONV risk factors BJA
and surgery-related risk factors have been associated with randomized controlled trials (RCTs) and large epidemiological
higher incidences of PONV. Although risk factors have merely observational studies that enrolled at least 500 adult
a correlative relationship with a given outcome, they can patients (age ≥15 yr old) and that identified independent
nevertheless be clinically useful. In contrast, independent pre- predictors of PONV by means of multivariate logistic regres-
dictors are more likely to have a causative relationship, and sion analysis. We included only studies in .500 patients
they may be used to predict or explain an outcome when stat- because large studies provide the most reliable data with a
istically corrected for other factors or confounders. Over the high level of evidence, which is in contrast to studies with
past few years, several groups have used multivariate logistic small sample sizes that can lead to randomly high (or low)
regression analyses to pinpoint which risk factors for PONV point estimates. Because of the many assumptions that are
are independent predictors. In some cases, scores based needed for multivariable analyses, there is no accurate stand-
on these independent predictors were developed, and in ard formula for sample size estimations available, but we
general, the attention paid to PONV has greatly improved in agreed that a sample size of 500 would provide a reasonable
clinical practice.2 4 5 12 – 14 threshold. Duplicate data, that is, studies reporting the same
While general reviews on PONV, including one review dedi- data in more than one publication, were excluded and only
cated to risk factors,11 reflect authors’ opinions, there has the data published in the primary article were included in the
743
BJA Apfel et al.
Table 1 Study characteristics. Twenty-two studies with a total of 95 154 patients were included. Median number of patients per study: 1505
(IQR: 1075– 2906). Overall occurrence of PV, PN, and PONV across all 22 studies: 28.25% (IQR: 18.9 – 39)
subpopulations with PN and PONV are very similar; thus, data in the meta-analysis, including variables that failed statistical
on PN and PONV were combined for statistical analyses. Quan- significance. If reported, the non-significant OR was used; non-
titative analyses were conducted for each risk factor that had significant ORs that were not reported were assigned an OR of 1
been previously identified in a multivariable analysis estimate (i.e. non-significant),16 since the 95% CI of a non-significant OR
in at least three studies. All variables that were assessed and includes 1. This approach was chosen because it is more conser-
analysed in the included study’s regression model were included vative than ignoring the evidence that a factor did not differ
744
Systematic review of PONV risk factors BJA
Table 2a. ORs of patient-, anaesthesia-, and surgery-related risk factors of PN, PV, and PONV as reported in the included studies. ORs that were
statistically significant in the original study are in bold; ORs that were not statistically significant are italicized. Risk factors that were found to be
non-significant in the original study but for which no OR was reported were included in the meta-analysis with an OR of ‘1’. Risk factors not
considered or reported in the study are labelled with ‘—’.
Study End Female History History History Non- Age BMI ASA Duration Volatile Nitrous Opioids- Opioids-
point gender of of of smoking (per (per anaes- oxide intra post
PONV MS migraine decade) hour) thetics operative operative
or MS
Cohen and N 2.6 – – – 1.8 0.9 b 1.0 j 1.5 m 1.5 p 1.5 u – 1.3 –
colleagues32
Koivuranta and N 2.4 2.3 a 1.7 1.6 2.1 1c 1.6 k 1.2 n 2.0 1.7 u – – 1.7
colleagues4 V 2.7 1.9 a 1.9 1 1.7 1.2 d 1.4 k 1.7 n 2.1 1u – – 1
Apfel and V 3.6 1.9 – – 2.1 0.8 1 – 1.3 – – 1 –
colleagues12
745
BJA Apfel et al.
N, postoperative nausea; V, postoperative vomiting; NV, postoperative nausea and/or vomiting; MS, motion sickness. aPONV only. b ≥70 vs ,50 yr; converted assuming
difference of 20 yr. c .50 vs ≤50 yr (n.s.). d .50 vs ≤50 yr; converted assuming difference of 20 yr. e ,50 vs ≥50 yr (n.s.). fper years; converted calculating (reported OR)10.
g
Children vs adults; not considered for including children data. h ,55 vs ≥55 yr; converted assuming difference of 20 yr. iPer 5 yr; converted calculating (reported OR)2.
j
Obesity/BMI.30. kBMI .25 vs ≤25. lper kg m22. mASA I –II vs ASA III– IV. nASA I vs II and I vs III combined. oASA I vs others. p60– 119 vs ,60 min. qOR per 30 min
converted by calculating OR2. rPer minute; converted by calculating OR60. sPer 30 min; converted by calculating OR2. t .100 vs ≤100 min; converted calculating OR0.6.
u
Anaesthesia—surgery duration per 10 min; converted calculating OR6. v2.5 –3.4 vs ,2.5 h. wPer 10 min; converted calculating OR6. xVolatile vs regional anaesthesia.
y
Volatile vs (MAC, RA, LA, ChrPB) combined. zVolatile vs propofol; derived from OR of 0.4 for propofol vs volatile. aaIso vs propofol. bbVolatile vs propofol. ccGeneral anaesthesia
(volatile and/or propofol) vs loco-regional. ddVolatile vs propofol; derived from OR of 0.69 for propofol vs volatile. eeDerived from OR of 0.81 for the substitution of nitrogen for
N2O. ffAny opioid vs none. ggPACU opioids n.s./PCA-based opioids significant. hhIncluding thyroid. iiMiddle ear+lower abdominal surgery combined. jjNon-D&C. kk90%
laparoscopic procedures. llHysterectomies only. mmStrabismus surgery only. nnLaparoscopic procedures. ooIncluding throat. ppGeneral surgery. qqHernia repair. rrLower
abdominal+middle ear surgery combined. ssGynaecological procedures only. ttCholecystectomies only.
746
Systematic review of PONV risk factors BJA
statistically significantly from 1 in a large study of .500 incompletely reporting data.26 27 One relevant article
patients. The 95% CI was calculated for the study’s overall end- meeting the inclusion criteria was retrieved via hand-
point incidences in the risk and reference groups.17 ORs and searching.28
their corresponding 95% CIs were calculated as the principal The final analysis included 22 studies reporting data from
measures of effect, with P,0.05 considered statistically signifi- a total of 95 154 patients, with a median number of patients
cant. Data were combined by means of a random effects model. per study of 1505 (inter-quartile range 1002 –3645)
Heterogeneity was assessed by I2 analysis, which (Table 1).2 – 4 6 – 8 12 – 14 21 28 – 38 Seventeen studies reported
describes the percentage of total variation across studies data on the overall occurrence of PONV. Of these, one
due to heterogeneity instead of chance.18 I 2 was calculated reported data on ‘severe’ PONV and another reported data
as I 2 ¼100%×(Q2df)/Q, where Q is Cochrane’s heterogen- for both PONV and PV.8 34 Two studies4 6 reported data for
eity statistic and df the degrees of freedom. As negative PN and PV separately and another two12 29 reported the
values of I 2 are converted to zero, I 2 values lie between 0% overall occurrence of PV as the sole study endpoint. In one
(no heterogeneity) and 100% (maximum heterogeneity). study, data reported for the secondary endpoint were consid-
While individual ORs for each risk factor would be expected ered because the primary endpoint included data from chil-
to show some variation between studies simply by chance, dren.8 In one of the articles, the data from two study
747
BJA Apfel et al.
Table 3 PN/PONV: combined estimates for patient-, anaesthesia-, and surgery-related predictors. For each risk factor, the number of studies in
which it was considered, the total number of patients, the ORs and respective 95% CIs, the degree of heterogeneity within the comparison, and
the number of outliers are given
Risk factors Studies (n) Participants (n) Combined estimate P-value Heterogeneity, I2 Outliers (n)
[OR (95% CI)]
Patient
Female gender 20 90 916 2.57 (2.32 –2.84) ,0.001 69 —
History of PONV or MS 16 44 216 2.09 (1.90 –2.29) ,0.001 54 —
Non-smoking 19 90 116 1.82 (1.68 –1.98) ,0.001 45 —
Age (per decade) 9 70 562 0.88 (0.84 –0.92) ,0.001 64 4
ASA 3 22 371 1.21 (0.88 –1.67) 0.24 86 0
BMI 4 20 428 1.00 (0.98 –1.02) 0.8 0 3
History of migraine 2 1778 1.77 (1.36 –2.31) — — —
(mean¼46)
the history of PONV or motion sickness (2.32) and non- any of the significant patient-specific (Supplementary Figs
smoking status (1.78) (Supplementary Table S1). Unlike for S6 and S7) or anaesthesia-related (Supplementary Fig. S8)
PN/PONV, age (in decades) was not an independent predictor risk factors.
of PV. The history of migraine, history of motion sickness,
high BMI, and low ASA physical status were not analysed
(n,3). None of the 13 surgical categories reached the level Discussion
of significance as an independent predictor of PV. This is the first systematic review and meta-analysis to iden-
For PN/PONV, heterogeneity was greatest in the surgery- tify and quantify the impact of independent predictors in
related risk factors (mean I 2 ¼70%), followed by a more mod- adults. Combining individual study results into the most ac-
erate degree of heterogeneity in the anaesthesia-related risk curate point estimate for each risk factor facilitated a direct
factors (mean I 2 ¼58%). The lowest overall heterogeneity comparison of the predictive strength of a comprehensive
was seen among the patient-specific predictors (mean list of predictors. The data suggest that PONV is mainly trig-
I 2 ¼46%; Table 3). gered by perioperative administration of emetogenic stimuli
Funnel plots indicated an underlying publication bias (volatile anaesthetics, prolonged duration of anaesthesia,
towards positive results in the significant surgical categories nitrous oxide, postoperative opioids) to susceptible patients
(Fig. 3). In contrast, there was no funnel plot distortion for (women, patients with a history of PONV and/or motion
748
Systematic review of PONV risk factors BJA
2.1.3 PV
Apfel 2002-PV 0.89 0.255 15.0% 2.44 (1.48, 4.01)
Apfel 1998A 1.28 0.24 16.2% 3.60 (2.25, 5.76)
Stadler 2003-PV 1.33 0.24 16.2% 3.78 (2.36, 6.05)
Apfel 1998B 0.55 0.227 17.4% 1.73 (1.11, 2.70)
Koivuranta 1997-PV 1 0.097 35.1% 2.72 (2.25, 3.29)
Subtotal (95% CI) 100.0% 2.73 (2.15, 3.46)
Heterogeneity: Tau2 = 0.03; Chi2 = 7.37, d.f. = 4 (P = 0.12); I 2 = 46%t
Test for overall effect: Z = 8.31 (P < 0.00001)
Fig 1 Forest plot showing ORs and 95% CIs for the risk factor female gender for the endpoints PN and/or PONV and PV using inverse variance in
a random effects model.
sickness, non-smokers, and patients of younger age). relatives with a history of PONV than those with no history
However, the evidence for a number of other factors which of PONV.43
are presumed to have an effect is either insufficient or The underlying mechanism for the reduced incidence of
lacking (e.g. menstrual cycle, type of surgery). Thus, inclusion PONV in smokers compared with non-smokers is also not
of these factors may compromise objective assessment of well understood. One theory suggests that chronic exposure
the patient’s risk for PONV. to polycyclic aromatic hydrocarbons in cigarette smoke might
Female gender was the strongest overall predictor of induce the cytochrome P450 isoenzymes (CYP2E1)44 45 re-
PONV with an OR of about 2.6. The incidence of PONV sponsible for phase 1 (first pass) metabolism of volatile anaes-
varies with the phase of the menstrual cycle,39 – 42 but men- thetics.46 47 However, given that only a small percentage of
strual hormonal fluctuations are unlikely to be responsible volatile anaesthetics gets metabolized (e.g. 0.2% of isoflurane,
for PONV. This has been confirmed in an RCT of .5000 0.02% of desflurane), it appears unlikely that liver enzyme in-
patients, which demonstrated no link between menstrual duction could account for such wide variation in the incidence
cycle phase or menopausal status and incidence of PONV.21 of PONV between smokers and non-smokers. Thus, we believe
The mechanism relating female gender to increased inci- that the protective effect of smoking may be due to functional
dence of PONV is as yet unknown. changes in neuroreceptors from chronic exposure to nicotine,
A history of PONV, motion sickness, or both indicates and thus nicotine withdrawal rather than nicotine exposure
an underlying susceptibility to PONV. A genome-wide associ- reduces smokers’ susceptibility to PONV.
ation study to identify potential genetic markers for suscep- The incidence of PONV generally decreases with age.
tibility to PONV found that patients with a history of severe, However, it should be noted that this is true for adults only,
intractable PONV were more likely to have first-degree as the incidence of POV increases with age in children, with
749
BJA Apfel et al.
3.1.7 Laparoscopic
Apfel 2004A 1.17 0.37 7.8% 3.22 [1.56, 6.65]
Apfel 1999 GER 0 0.29 10.4% 1.00 [0.57, 1.77]
Cohen 1994 0.83 0.28 10.7% 2.29 [1.32, 3.97]
Apfel 2002 0.75 0.28 10.7% 2.12 [1.22, 3.66]
Apfel 1999 FIN 0 0.23 12.9% 1.00 [0.64, 1.57]
Koivuranta 1997A 0 0.21 13.8% 1.00 [0.66, 1.51]
Bosch 2005 0 0.2 14.3% 1.00 [0.68, 1.48]
Choi 2005 0.29 0.1 19.4% 1.34 [1.10, 1.63]
Subtotal (95% CI) 100.0% 1.37 [1.07, 1.77]
Heterogeneity: Tau² = 0.08; Chi² = 18.61, df = 7 (P = 0.009); I² = 62%
Test for overall effect: Z = 2.47 (P = 0.01)
Fig 2 Forest plot showing ORs and 95% CIs for the risk factors of gynaecological surgery, cholecystectomy, and laparoscopic procedures for
the endpoint PN and/or PONV using inverse variance in a random effects model.
a relatively low reported incidence below the age of 3.44 – 46 Opioid analgesia primarily involves central m, k, and d
An underlying mechanism may be reduced autonomic receptors in the rostral anterior cingulate cortex and the
reflexes with increasing age. brainstem.51 However, opioid activity at peripheral receptors
Of the anaesthesia-related risk factors, the use of volatile in the gut inhibits the release of acetylcholine from the mes-
anaesthetics was the strongest predictor, followed by the enteric plexus and stimulates m receptors, which reduces
duration of anaesthesia, postoperative opioid use, and muscle tone and peristaltic activity. Consequent delayed
nitrous oxide. In a study of 1180 patients, volatile anaes- gastric emptying and gastric distension activate visceral
thetics were the single greatest factor affecting the incidence mechanoreceptors and chemoreceptors, which trigger
of emesis in the first 2 h after operation, and volatile anaes- nausea and vomiting via a serotonergic signalling pathway.
thetic use increased PONV in a dose-dependent manner irre- Whether certain types of surgery are associated with a
spective of the choice of the agent.8 21 Nitrous oxide may higher incidence of PONV has been controversial. In our ana-
contribute to PONV in several ways. Nitrous oxide may act lysis, only cholecystectomy, laparoscopic procedures, and
upon the dopamine47 and opioid48 receptors in the brain, gynaecological surgery reached statistical significance as
produce changes in middle ear pressure,49 50 and/or cause independent predictors of PONV, and no type of surgery
bowel distension as it diffuses into closed cavities.48 reached statistical significance as a predictor of PV.
750
Systematic review of PONV risk factors BJA
SE(log[OR])
0
0.1
0.2
0.3
Subgroups
0.4 Gynaecological procedures
Cholecystectomy
Laparoscopic procedures
Fig 3 Funnel plot of the effect of gynaecological surgery, cholecystectomy, and laparoscopic procedures on PN and/or PONV endpoint.
However, it must be noted that surgery reference groups dif- Supplementary material
fered widely between studies, which may have led to a bias
Supplementary material is available at British Journal
towards positive results. Consequently, to what extent
of Anaesthesia online.
surgery-related risk factors are truly independent predictors
remains unclear due to the potential for attribution error
and presence of heterogeneity.11 52 In contrast, a lower
degree of heterogeneity and a lower risk of bias were Acknowledgement
observed for patient-specific and anaesthesia-related com- The authors would like to thank Gloria Y. Won, MLIS, for
parisons, which makes them more reliable independent providing guidance during our literature search.
predictors.
Of the six risk scores11 for the prediction of PONV in adult
surgical patients, only two13 14 consider surgery type to be an Declaration of interest
independent predictor, while the majority2 4 5 12 do not. Two
None declared.
validated and widely used scores are the score developed by
Koivuranta and colleagues4 and Apfel and colleagues;2 the
latter is based on a cross-validation of Koivuranta’s and the
author’s own data. The cross-validation showed that the
Funding
duration of anaesthesia was highly correlated with post- This work was supported by Dr Apfel’s Perioperative Clinical
operative opioid use, and therefore did not add to the pre- Research Core.
dictive power of the overall score.
Based on our systematic review and meta-analysis, there
is strong evidence that several patient-specific and References
anaesthesia-related characteristics are the strongest inde- 1 Dolin SJ, Cashman JN, Bland JM. Effectiveness of acute post-
pendent risk factors for PONV. Evidence for the type of operative pain management: I. Evidence from published data.
surgery is at best conflicting and could be biased as a Br J Anaesth 2002; 89: 409– 23
result of inconsistent reference groups. Including factors 2 Apfel CC, Laara E, Koivuranta M, Greim CA, Roewer N. A simplified
with a limited evidence base may be counterproductive risk score for predicting postoperative nausea and vomiting:
conclusions from cross-validations between two centers.
when assessing a patient’s risk of PONV. The simplified
Anesthesiology 1999; 91: 693– 700
Apfel score (female gender, history of PONV or motion
3 Eberhart LHJ, Hogel J, Seeling W, Staack AM, Geldner G,
sickness, non-smoking status, postoperative opioids) and
Georgieff M. Evaluation of three risk scores to predict post-
Koivuranta’s risk score (female gender, history of PONV operative nausea and vomiting. Acta Anaesthesiol Scand 2000;
or motion sickness, non-smoking status, age, duration of 44: 480– 8
surgery) are logical choices for an objective PONV risk assess- 4 Koivuranta M, Laara E, Snare L, Alahuhta S. A survey of post-
ment in daily clinical practice. operative nausea and vomiting. Anaesthesia 1997; 52: 443–9
751
BJA Apfel et al.
5 Palazzo M, Evans R. Logistic regression analysis of fixed patient 26 Chia YY, Kuo MC, Liu K, Sun GC, Hsieh SW, Chow LH. Does post-
factors for postoperative sickness: a model for risk assessment. operative pain induce emesis? Clin J Pain 2002; 18: 317–23
Br J Anaesth 1993; 70: 135– 40 27 Kredel M, Goepfert C, Bassi D, Roewer N, Apfel CC. The influence
6 Stadler M, Bardiau F, Seidel L, Albert A, Boogaerts JG. Difference in of the weather and the phase of the moon on post-
risk factors for postoperative nausea and vomiting. Anesthesi- operative nausea and vomiting. Acta Anaesthesiol Scand 2006;
ology 2003; 98: 46– 52 50: 488–94
7 Visser K, Hassink EA, Bonsel GJ, Moen J, Kalkman CJ. Randomized 28 Wallenborn J, Gelbrich G, Bulst D, et al. Prevention of post-
controlled trial of total intravenous anesthesia with propofol operative nausea and vomiting by metoclopramide combined
versus inhalation anesthesia with isoflurane-nitrous oxide: post- with dexamethasone: randomised double blind multicentre
operative nausea with vomiting and economic analysis. trial. Br Med J 2006; 333: 324
Anesthesiology 2001; 95: 616– 26 29 Apfel CC, Greim CA, Haubitz I, et al. The discriminating power of
8 Apfel CC, Kranke P, Katz MH, et al. Volatile anaesthetics may be a risk score for postoperative vomiting in adults undergoing
the main cause of early but not delayed postoperative vomiting: various types of surgery. Acta Anaesthesiol Scand 1998; 42:
a randomized controlled trial of factorial design. Br J Anaesth 502– 9
2002; 88: 659– 68 30 Apfel CC, Kranke P, Eberhart LHJ, Roos IA, Roewer N. A compari-
9 Scuderi PE, Conlay LA. Postoperative nausea and vomiting and son of predicting models for postoperative nausea and vomiting.
752
Systematic review of PONV risk factors BJA
candidate markers mediating susceptibility to postoperative postoperative nausea and vomiting. A meta-analysis. Anesthesi-
nausea and vomiting. Anesthesiology 2011; 115: 54–64 ology 1996; 85: 1055– 62
44 Kretz FJ. Postoperatives Erbrechen: Medikamentoese Prophylaxe 49 Nader ND, Simpson G, Reedy RL. Middle ear pressure changes
und Therapie mit etablierten Substanzen beim Kind. Anasthesiol after nitrous oxide anesthesia and its effect on postoperative
Intensivmed Notfallmed Schmerzther 1997; 32: 620– 2 nausea and vomiting. Laryngoscope 2004; 114: 883– 6
45 Rose JB, Watcha MF. Postoperative nausea and vomiting in paedi- 50 Chinn K, Brown OE, Manning SC. Effects of inhalant anesthesia on
atric patients. Br J Anaesth 1999; 83: 104– 17 the middle ear as measured by tympanometry. Arch Otolaryngol
46 Apfel CC, Stoecklein K, Lipfert P. PONV: a problem of inhalational Head Neck Surg 1993; 119: 283– 7
anaesthesia? Best Pract Res Clin Anaesthesiol 2005; 19: 51 Machelska H, Stein C. Immune mechanisms in pain control.
485 – 500 Anesth Analg 2002; 95: 1002–8
47 Murakawa M, Adachi T, Nakao S, Seo N, Shingu K, Mori K. Activa- 52 Watcha MF, White PF. Postoperative nausea and vomiting. Its
tion of the cortical and medullary dopaminergic systems by etiology, treatment, and prevention. Anesthesiology 1992; 77:
nitrous oxide in rats—a possible neurochemical basis for psycho- 162– 84
tropic effects and postanesthetic nausea and vomiting. Anesth 53 Apfel CC, Kranke P, Eberhart LH. Comparison of surgical site and
Analg 1994; 78: 376– 81 patient’s history with a simplified risk score for the prediction of
48 Divatia JV, Vaidya JS, Badwe RA, Hawaldar RW. Omission of postoperative nausea and vomiting. Anaesthesia 2004; 59:
753